Avatar for Username

Pomoc pśepytaś

Glědajśo se wobšudy pomocy. Njenapominajomy was nigda, telefonowy numer zawołaś, SMS pósłaś abo wósobinske informacije pśeraźiś. Pšosym dajśo suspektnu aktiwitu z pomocu nastajenja „Znjewužywanje k wěsći daś“ k wěsći.

Dalšne informacije

Toś ta nitka jo se archiwěrowała. Pšosym stajśo nowe pšašanje, joli trjebaśo pomoc.

HTTP request (GET) executed twice after RST TCP

  • 3 wótegrona
  • 1 ma toś ten problem
  • 10 naglědow
  • Slědne wótegrono wót afcinza

afcinza
afcinza
more options

I have a load balancer that is executing a reset in the TCP connections when I try to execute a PHP script which takes a long time to finish (400 sec). We are debugging this but the problem is that, when Firefox receives the RST, it executes the same query again. This can generate big problems. You can see the problem in the next Wireshark screenshot. Any idea about why Firefox is doing this?

Thank you in advance. Regards.

I have a load balancer that is executing a reset in the TCP connections when I try to execute a PHP script which takes a long time to finish (400 sec). We are debugging this but the problem is that, when Firefox receives the RST, it executes the same query again. This can generate big problems. You can see the problem in the next Wireshark screenshot. Any idea about why Firefox is doing this? Thank you in advance. Regards.

Wšykne wótegrona (3)

afcinza
afcinza Pšašaŕ
more options

I attach the Wireshark screenshot. Thanks!

cor-el
cor-el
  • Moderator
  • Top 10 Contributor
more options

See:

HTTP connections now have a response timeout

If your application is processing large data and affected by this change, you can change the value of the network.http.response.timeout preference to a larger number via about:config.

  • Bug 947391 – HTTP connections (exc. XHR, SPDY) should have a response timeout
afcinza
afcinza Pšašaŕ
more options

Thank you for your quick answer, cor-el. I knew about this parameter but the main problem is the second GET that Firefox executes automatically after receiving the TCP RST. The browser should not execute a second GET without user control, shouldn't it? This is strange but this behaviour is reproducible.

The PHP scripts execute some changes into a database. If the browser executes the HTTP request without user control this may produce an unexpected result.

Thank you!